Predictors of preoperative MRI for breast cancer: differences by data source

Abstract

Aim Investigate how the results of predictive models of preoperative magnetic resonance imaging (MRI) for breast cancer change based on available data Materials and Methods 1919 insured women aged ≥18 with stage 0-III breast cancer diagnosed 2002-2009. Four models were compared using nested multivariable logistic, backwards stepwise regression; model fit was assessed via area under the curve (AUC), R2. Results MRI recipients (n=245) were more recently diagnosed, younger, less comorbid, with higher stage disease. Significant variables included: Model 1/Claims (AUC=.76, R2=0.10): year, age, location, income; Model 2/Cancer Registry (AUC=.78, R2 =0.12): stage, breast density, imaging indication; Model 3/Medical Record (AUC=.80, R2 =0.13): radiologic recommendations; Model 4/Risk Factor Survey (AUC=.81, R2 =0.14): procedure count. Conclusions Clinical variables accounted for little of the observed variability compared to claims data.

Preoperative Breast Magnetic Resonance Imaging Use by Breast Density and Family History of Breast Cancer

“…The prior study by Loggers et al found that women with nondense breasts were 48% less likely to have MRI compared with women with dense breasts. 27 This finding is in contrast to our result of no association and may be due to differences in the study population ages (we included 66 and older whereas Loggers et al included 18 and older), geographic areas, adoption of clinical standards, and healthcare system types (we included all women in BCSC registries and Loggers et al included one HMO in the northwestern United States), or covariates ad-justed for in the models. In another study, 1012 breast surgeons in the United States were surveyed in December 2010 and asked if they routinely recommended breast MRI to women newly diagnosed with breast cancer (routinely was defined as ‡75% of the time).…”